ORIGINAL ARTICLE
Estimating drinking-water ingestion and dermal contact with water in a French population of pregnant women: the EDDS cohort study
Marion Albouy-Llaty1,2, Antoine Dupuis1,2, Claire Grignon1,2, Sylvie Strezlec3, Fabrice Pierre4, Sylvie Rabouan1and Virginie Migeot1,2
The aim of the present study, a part of the Endocrine Disruptor Deux-Se`vres (EDDS) cohort study, was to estimate water-use habits of pregnant French women. The study population consisted of 132 pregnant women living in Deux-Se`vres (France) in 2012–2013, in areas where drinking water is exclusively produced by surface water. Drinking-water data included ingested water (tap, bottled and filtered) and ingestion place (home, work and elsewhere). Dermal contact with water included showering, bathing, swimming, spa use, hand-washing and other water activities. Data were collected through face-to-face interviews at second and third trimesters of pregnancy with a 1-day-recall questionnaire. Intertrimestral differences in water-use habits were assessed. Predictors of water ingestion and duration of dermal contact with water were assessed with multiple linear regressions. At the second trimester of pregnancy, the mean total drinking-water ingestion was 1.8±0.6 l per day (mean and SD), 71% of which was tap water.
Total drinking-water ingestion was not different between both trimesters but ingestion place differed. Dermal contact with water estimate was 188±118 and 173±92 min/week at second and third trimesters, respectively. Smoking increased water ingestion 777 ml/day 95% CI (171–1384). Duration of dermal contact in spring was 30 min/week 95% CI (13–48) higher than in winter. Obese women spend 26 min/week 95% CI (2–50) more showering than women with recommended weight. Our estimates of pregnant French women’s exposure to water will help researchers to better assess water pollutant risks.
Journal of Exposure Science and Environmental Epidemiology(2015)25,308–316; doi:10.1038/jes.2014.48; published online 30 July 2014 Keywords: water contaminants; drinking-water ingestion; dermal contact with water; pregnancy; France
INTRODUCTION
In assessment of water pollutant risk, water-use habits of pregnant women have been estimated by questionnaires or diaries, in Italy, Spain, USA and Great Britain.1–10Some studies have also assessed predictors of water-use habits during pregnancy.4,5,9,10 Water ingestion is contingent on geographical and cultural factors;1so it is necessary to adapt data to the local context. Only one study has assessed water-use habits of French pregnant women.11 This study assessed water ingestion at early pregnancy and described variations in consumption during pregnancy.5 Furthermore, it assessed showering, bathing, swimming habits at 2-year follow up after birth, which could entail recall bias.11
There are few birth cohorts that are monitored for studying water contaminants.12 They are centered on water disinfection by-products, although other chemical water contaminants also exist.13,14 The Endocrine Disruptor Deux-Se`vres (EDDS) cohort study was performed in French pregnant women to estimate exposure to endocrine disruptors with different methods from science exposure: environmental monitoring, biomonitoring in breast milk and questionnaire.15
We present here the results of a study originating in the EDDS cohort of which the aim was to accurately estimate drinking-water ingestion and dermal contact with water in a French population of pregnant women. The secondary objectives were to estimate
difference in water-use habits between second and third trimester of pregnancy and to determine predictors of the latter.
METHODS Population
Eligible subjects were pregnant women who declared their pregnancy before 14 weeks of gestation, in 2012–2013, and were living in one of the 87 Deux-Se`vres municipalities. In these municipalities, drinking-water is exclusively produced from surface water because EDDS cohort objectives focused on emerging chemical contaminants that were more frequently found in surface water than groundwater, in 201116,17(Supplementary File 1).
Exclusion criteria were (i) women who did not wish to breastfeed their neonate, (ii) women who decided to deliver outside the five maternities enrolled in the study, (iii) women not living in one of the 87 municipalities of the Deux-Se`vres district 1 year before pregnancy or who would move during pregnancy, (iv) maternal ageo18 years (v) women deprived of their liberty following a legal or administrative decision, hospitalized without their consent, admitted to a public health or social institution, aged418 years but subject to a legal protection order and/or unable to express their consent, (vi) multiple pregnancy, (vii) women who did not speak French and (viii) women not affiliated to social insurance (Figure 1).
The study was approved by the local ethical committee and declared to the French independent administrative data protection authority: ‘‘Com- mission Nationale de l’Informatique et des Liberte´s’’.
1Department of Analytical Chemistry, Pharmaceutics and Epidemiology, Faculty of Medicine and Pharmacy, University of Poitiers, IC2MP, UMR7285-CNRS, Poitiers Cedex France;
2Poitiers University Hospital, Biology-Pharmacy-Public Health Pole, Poitiers Cedex, France;3Maternal and Childhood Protection, Conseil general des Deux-Se`vres, Niort, France and4Poitiers University Hospital, Mother and child Pole, Poitiers Cedex, France. Correspondence: Dr. Marion Albouy-Llaty, Department of Analytical Chemistry, Pharmaceutics and Epidemiology, Faculty of Medicine and Pharmacy, University of Poitiers, IC2MP, UMR7285-CNRS, 6 rue de la Miletrie, Poitiers Cedex 86034, France.
Tel.:þ33 549454326.
E-mail: marion.llaty@univ-poitiers.fr
Received 5 December 2013; revised 20 May 2014; accepted 20 May 2014; published online 30 July 2014
Recruitment and Data Collection
The women were recruited following compulsory declaration of pregnancy sent to Maternal and Childhood Protection, the ‘‘protection maternelle et infantile’’ (PMI) ward. From declaration of pregnancy, we collected presume date of delivery, address, date and place of mother birth, number of previous pregnancies and employment.
Then, a targeted mail was sent to eligible women for age and municipality. When they did not answer, a reminder letter was sent;
moreover, during routine visits PMI midwives explained the study to them.
Once the women had acknowledged receipt, a researcher (CG) called to set up an appointment at their homes. Written consent signed by both future parents was obtained at first visit.
The pregnant women received 1-h visits by a trained interviewer during the second and the third trimesters of pregnancy. Within the realms of possibility, we scheduled visits to address possible day of week and seasonal effects. Data collection procedures are shown in Supplementary File 2. The first and the second prenatal questionnaires were administered to the women during the home visits. The birth questionnaire was filled in by a maternity midwife in the first 72 h after delivery.
The questionnaires were expressly designed for the EDDS study (Supplementary File 3). They comprised questions chosen following an intensive review of the literature on water-use habits (Supplementary File 4) and exposure assessment questionnaire design.18 The questionnaires had five sections: (i) socioeconomic data,19(ii) medical data, (iii) history of
professional exposure, (iv) dermal water exposure,4,5,8–10(v) drinking-water and bottled water ingestion frequency,4–6,8–10(vi) diet questions (rate of cans, fresh fruit and vegetable consumption), (vii) alcohol and tobacco consumption and (viii) cosmetics type and use frequency.
Water Ingestion Estimation
Drinking-water ingestion data were estimated with a 1-day hour-by-hour recall questionnaire for a weekday (the day before interview) and a weekend day (the most recent).
Ingestion was reported in number of different containers with a picture:
100 ml for a cup of coffee, 150 ml for a medium glass, 150 ml for a plastic glass, 200 ml for a large glass, 250 ml for a mug and 400 ml for a bowl with frequency of use (Supplementary File 5).
Content was assessed by type of water: tap water, coffee, tea or other beverages, such as reconstituted fruit juices; filtered tap, boiled tap or bottled water. Finally, the ingestion place was recorded: home, work and elsewhere. Moreover, women were asked whether they had modified their ingestion since the beginning of the pregnancy and whether their home tap water was treated (softened and filtered).
We calculated the total water ingestion during a week by adding all types of water ingestion. We attributed a 5-weight to the volume consumed in a weekday and a 2-weight to the volume consumed in a weekend day. Then, we divided by 7 to obtain average daily consumption.
Figure 1. Flow chart of the EDDS cohort study, Deux-Se`vres on February 2014.
309
We calculated the proportion of total water ingested that comes from tap water as volume of tap water and beverages produced from tap water, divided by volume of total water ingested. We calculated proportion of total tap water ingested that comes from home tap water as volume of home tap water divided by volume of total tap water (home, work and elsewhere). We calculated proportion of home tap water that comes from filtered home tap water as volume of filtered home tap water divided by volume of home tap water.
Dermal contact with water data was estimated with frequency and duration of showering, bathing, swimming, using spa or Jacuzzi, hand- washing, dish-washing, child’s bath, pets’ bath and other water activities during a week. It was expressed in minutes per week. We calculated proportion of total dermal water that comes from each water activity (showering, bathing, swimming, hand-washing and others) as duration of each activity per week divided by duration of total dermal water contact per week.
We calculated the average percentage of each proportion.
Socioeconomic Data
Individual socioeconomic status was assessed by the EPICES (Evaluation de la Pre´carite´ et des Ine´galite´s de sante´ dans les Centres d’Examens de Sante´) score. It takes into account the multidimensional characteristics of socioeconomic status. It is a score ranging from 0 (least deprived) to 100 (most deprived) composed of 11 items: contact with a social worker, complimentary insurance, couple, homeowner, difficulties meeting funda- mental needs, practising sport, going to shows, holidays, relations with family and friends and having a person to help if necessary. It is associated with health status independently of occupational category.19
Neighborhood socioeconomic status was characterized by the census- based Townsend index.20 The geographical units used were ‘‘ıˆlots regroupe´s pour l’information statistique’’ (IRIS) or regrouped statistical information blocks, as defined by the French National Institute for Statistics and Economic Studies (INSEE), an IRIS representing the smallest geo- graphical census unit available in France. Major towns are divided into several IRIS units and small towns form a single IRIS. Each IRIS includes B2000 individuals with relatively homogeneous social characteristics. The Townsend index is the sum of four aggregate variables related to socio- economic context of IRIS (unemployment, non-car ownership, non-home ownership and household overcrowding). The higher the index the higher the IRIS is considered to be deprived. The data needed for its construction were drawn from the population census carried out in 2007 by the INSEE.
In 2007, Deux-Se`vres counted 305 municipalities and 362 IRIS. Mothers’
addresses have been geocoded at the IRIS level through a correlation map effective in 1999,21 provided by the Maurice Halbwachs Center, which collects surveys and databases following agreements with the INSEE, several ministerial statistical services and other public institutions.
The Townsend index was categorized in tertiles based on the population studied.
Individual Data
Maternal age was divided into two categories according to the median: 31 and less and over 31 years of age. Parity was defined as number of children: none, one and two or more. Smoking during pregnancy was categorized as non-smoker and current smoker with number of cigarettes per day. Alcohol consumption was categorized as non-consumer and current consumer. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared (kg/m2) and categorized as underweight (o18.5), recommended weight (18.5–24.9), overweight (25.0–29.9) or obese (Z30.0). Marital status was defined in two categories: in couple (married or not) or single (widow, never married or divorced).
Education level was defined as the highest educational qualification in four categories (primary and secondary school, French baccalaureate, baccalaureateþ2 years and higher education). Employment was defined as women’s occupation at time of questionnaire (yes, no). Household income was divided into seven categories: 451–800h/month, 801–1500h/
month, 1501–2300h/month, 2301–3000h/month, 3001–3800h/month, 3801–4500h/month and44500h/month.
Statistical Analysis
Consumption of tap and bottled water was estimated as was dermal contact with water from bathing, showering and hand-washing. Quanti- tative variables were expressed as mean, SD, minimum, maximum and
percentiles for women who had been exposed by this pathway. Counts of qualitative variables were expressed as frequency and percentages.
We examined differences in drinking-water ingestion for all women between trimesters of pregnancy by a paired Wilcoxon test taking the number of tests into account with Bonferroni correction.
We carried out multiple linear regressions with aim to identify predictors of tap water ingestion as an indicator of water ingestion and shower duration as an indicator of duration of water contact. We considered variables that could be associated with water-use habits: season, age, parity, socioeconomic data and healthy behaviors, such as nutrition (or BMI) and smoking.
The analyses were performed using SAS version 9.3 (Cary, NC, USA).
RESULTS Study Population
In this study, 132 women were included. These women lived in one of the 87 eligible municipalities in which Townsend score distribution was not different than that in non-eligible munici- palities (data available from the authors).
Table 1. Comparison between pregnant women who did not participate and those who participated in the EDDS Cohort study, Deux-Se`vres, France, 2012–2013.
Women who did not participate
Women who participated
N¼1238 N¼166
N % N % P
Contextual data Area
Bocage 490 39.6 63 38.0 0.465
Thoursais 75 6.1 6 3.6
Gaˆtine 315 25.4 42 25.3
Haut-Val-de-Se`vre 250 20.2 42 25.3
Mellois 108 8.7 13 7.8
Water community system
Gaˆtine Ce´ bron 195 15.8 24 14.5 0.785
Lambon Fressines 86 7.0 10 6.0
Lambon La Cre`che 50 4.0 9 5.4
Saint Maixent re´ gion 183 14.8 29 17.5 Thorigne´ Saint Le´ ger 39 3.2 7 4.2 Val de Loire Ce´ bron 685 55.3 87 52.4 Townsend scorea
Tercile1: least deprived 371 32.2 55 35.0 0.456
Tercile2 340 29.5 50 31.9
Tercile3: most deprived 441 38.3 52 33.1 Individual data
Maternal age (years)a
(18–25) 222 17.9 9 5.4 o103
(25–30) 502 40.6 59 35.5
(30–35) 357 28.8 63 38.0
Z35 157 12.7 35 21.1 Place of birtha
France 1149 94.6 157 95.2 0.755
Elsewhere 66 5.4 8 4.8
Employmenta
No 306 25.8 26 16.0 0.006
yes 879 74.2 137 84.0
Number of previous pregnanciesa
0 433 38.5 49 30.8 0.007
1 428 38.0 55 34.6
2 176 15.6 31 19.5
Z3 89 7.9 24 15.1
aMissing data.
Bold values are statistically significant relation at 5%.
310
The flow chart is presented in Figure 1. The mean delay between both visits was 80±22 days (41–133) and 107 (81%) pregnant women were visited at two different seasons. Moreover, 102 (77%) pregnant women were visited at two different days of week.
Differences between participants and non-participants are pre- sented Table 1. The participant women were older, had more pre- vious pregnancies and were more employed than non-participant women.
Sociodemographic characteristics of study population are pre- sented in Table 2. On the average, women were 31.3±4.2 (21–45) years of age.
Water-use Habits
Estimates of drinking-water ingestion and duration of dermal contact with water are presented in Table 3. In all places of inge- stion, 124 (94%) women declared having consumed tap water, on average 1.4±0.8 l/day at the second trimester (1.3±0.7 l/day during winter and 1.8±0.9 l/day during spring), and 119 (90%) on average 1.5±0.9 l/day at the third trimester (1.4±0.9 l/day during winter and 1.5±0.8 l/day during spring). About half of women declared having consumed bottled water during pregnancy.
Duration of total dermal contact was, on average, 188±118 and 173±92 min/week, respectively, at second and third trimesters of pregnancy, that is to say 3 h.
The principal sources of water-use habits are presented in Table 3 as proportions.
Differences of Water-Use Habits Between Trimesters
Differences of water-use habits between second and third trimesters are presented in Table 3. According to Bonferroni correction, there were no differences in total drinking-water ingestion between both trimesters; however, we have observed an increase in tap water ingestion at home and a decrease in tap water ingestion at work. There was no difference of dermal Table 2. Sociodemographic characteristics of 132 pregnant women
in the EDDS Cohort who were visited twice during pregnancy, Deux-Se`vres, France, 2012–2013.
N %
Contextual data Area
Bocage 54 40.9
Thoursais 3 2.3
Gaˆtine 36 27.2
Haut-Val-de Se`vre 29 22.0
Mellois 10 7.6
Water community system
Gaˆtine Ce´ bron 21 15.9
Lambon Fressines 7 5.3
Lambon La Cre`che 8 6.0
Saint Maixent re´ gion 18 13.6
Thorigne´ Saint Le´ ger 6 4.6
Val de Loire Ce´ bron 72 54.6
Townsend scorea
Tercile1: IRIS least deprived 46 35.7
Tercile2 41 31.8
Tercile3: IRIS most deprived 42 32.5
Individual data Maternal age
r25 years 5 3.8
(25–30) 41 31.1
(30–35) 59 44.6
Z35 27 20.5
Marital status
Single 1 0.8
In couple 131 99.2
Education level
Primary, secondary school 23 17.4
French baccalaureate 21 16.0
Bacþ2 37 28.0
Higher education 51 38.6
Employment
No 22 16.7
Yes 110 83.3
BMI
Underweight:o18.5 17 12.9
Recommended weight (18.5–24.9) 78 59.1
Overweight (25.0–29.9) 25 18.9
Obese: (Z30.0) 12 9.1
Parity
0 42 31.8
1 53 40.2
Z2 37 28.0
Smokinga
Before pregnancy 31 23.5
Mean and SD number of cigarettes/
(min–max)
10.1±8.6 (1–40) At the first trimester of pregnancy 21 15.9 Mean and SD number of cigarettes/
(min–max)
7.0±7.6 (1–35)
At the second trimester 14 10.6
Mean and SD number of cigarettes/
(min–max)
7.9±6.4 (1–25)
At the third trimester 12 9.1
Mean and SD number of cigarettes/
(min–max)
7.9±5.3 (3–20) Alcohol consumption
At the first trimester 78 59.1
At the second trimester 15 11.4
Socio-professional status
Directors, executives, managers 3 2.7
Intellectual and scientific professionals 35 31.3
Intermediate professionals 28 25.0
Administrative employees 7 6.4
Merchants, traders and direct service providers
28 25.0
Table 2. (Continued).
N %
Farmers and skilled labor in agriculture, forestry and fishery
2 1.8
Skilled crafts and industrial professionals
2 1.8
Plant and machinery operators and assemblers
2 1.8
Unskilled professionals 3 2.7
Military professionals 2 1.8
Household incomeb
451–800h/month 1 0.8
801–1500h/month 4 3.0
1501–2300h/month 25 18.9
2301–3000h/month 41 31.1
3001–3800h/month 42 31.8
3801–4500h/month 12 9.1
44500h/month 7 5.3 EPICES scorec
1 (0–7.1; less deprived) 73 55.3
2 (7.1–16.6) 40 30.3
3 (16.6–30.2) 11 8.3
4 (30.2–48.5) 6 4.6
5 (48.5–100; most deprived) 2 1.5
aMissing data.
b1 Euro¼1.3645 US dollars.
cEPICES score is a French individual deprivation score.
311
contact with water duration between both trimesters except for hand-washing.
Factors Associated with Water-use Habits
Results of regression analysis of tap water ingestion and showering duration are presented in Tables 4 and 5. Smoking during the second trimester of pregnancy increased tap water ingestion 777 ml/day 95% CI (171;1384). Spending the second trimester of pregnancy during spring, compared with winter, increased shower duration about 30 min/week 95% CI (13;48).
Being obese increased shower duration about 26 min/week 95%
CI (2;50).
Diet Exposure
Among the women enrolled in the EDDS study and visited twice, 110 (83%) declared consumption of food packaged in cans and 13 (10%) declared consumption of beverages packaged in cans.
Moreover, 62 (47%) declared that they never eat organic fruits and vegetables.
DISCUSSION
Our study is the first one to assess water-use habits of French pregnant women twice during pregnancy. The French pregnant
women enrolled in the EDDS study consumed 1.8 and 1.9 l/day of total drinking water, respectively, at second and third trimesters, and were in contact with water 3 h/week.
The study was carried out in Deux-Se`vres district, which is more rural and less deprived than other districts of France. The rural residence and the socioeconomic status have been described as associated with water-use habits during pregnancy;3,4,9 therefore, this could have entailed a selection bias. However, cultural settings of women living in Deux-Se`vres may remain closer to French pregnant women than those of foreign women.
Despite the availability of water-use habit data in foreign countries,1–8our estimates seem more tailored-made for France.
Eligible women had to live in one of the 87 municipalities of the Deux-Se`vres district, whose socioeconomic context was no different than that of non-eligible municipalities. Eligible women had to breastfeed their neonates, although breastfeeding concerns only 56% of women living in Deux-Se`vres district.22 This criterion could have entailed a selection bias if breastfeeding is associated with water ingestion. We can hypothesize that breastfeeding may be associated with healthy behaviors, and bottled water ingestion is one of them.23On the contrary, we can hypothesize that breastfeeding may be associated with behaviors more protective for the planet, and tap water ingestion is one of them.23Moreover, EDDS women were perhaps more interested in this study than those who do not because they drink more tap Table 3. Drinking-water daily ingestion (ml/day) and dermal contact with water (minutes/week), in second and third trimesters of pregnancy, and their differences, EDDS Cohort Study, France, 2012–2013 (n¼132 pregnant women).
Second trimester Third trimester Pa
Volume/duration Mean (SD) proportion
Volume/duration Mean (SD) proportion
n Mean SD Min–max n Mean SD Min–max
Total drinking-water ingestion (ml/day)
132 1827 638 800–4650 132 1937 707 800–4850 0.029
Tap water
All place of intake 124 1415 809 71–3650 71 (35)b 119 1526 905 43–4100 70 (39)b 0.375
Filtered tap water 23 1301 796 71–3150 21 1642 728 464–3300
Boiled tap water 96 491 320 29–1900 81 486 376 43–2420
Intake at home 121 1313 807 71–3650 91 (17)c 119 1518 904 43–4100 99 (3)c 0.004a
Filtered tap water 21 1388 778 71–3150 83 (28)d 21 1642 728 464–3300 96 (14)d
Boiled tap water 91 478 318 29–1900 83 493 370 43–2420
Intake at work 40 404 305 71–1214 3 238 230 71–500 o103a
Filtered tap water 2 393 101 321–464 0
Boiled tap water 17 208 126 71–464 3 143 71 71–214
Bottled water 63 1042 561 150–2743 61 1215 671 43–3000 0.159
Total dermal contact (min/week)
132 188 118 55–1135 132 173 92 35–637 0.119
Showering 131 73 38 20–210 42 (16)e 132 75 43 14–280 46 (19)e 0.386
Bathing 42 33 34 4–158 16 (13)f 45 34 29 5–120 17 (13)f 0.451
Swimming at public facilities
23 59 30 15–180 26 54 19 23–120
Swimming at home 10 37 39 1–120 5 115 94 20–210
Spa 6 9 5 4–15 6 12 9 1–23
Hand-washing 132 75 77 12–630 38 (18)g 132 56 65 5–525 31 (17)g o103a
Other water activities 127 17 15 2–90 10 (8)h 126 16 14 1–60 11 (9)h 0.178
aWilcoxon signed rank sum test: according to Bonferroni correction,Po0.005 is significant.
bThe column mean (SD) proportion represents the proportion of total water ingested that comes from tap water.
cThe column mean (SD) proportion represents the proportion of total tap water ingested that comes from home tap water.
dThe column mean (SD) proportion represents the proportion of home tap water that comes from filtered home tap water.
eThe column mean (SD) proportion represents the proportion of dermal contact with water duration that comes from shower.
fThe column mean (SD) proportion represents the proportion of dermal contact with water duration that comes from bath.
gThe column mean (SD) proportion represents the proportion of dermal contact with water duration that comes from hand-washing.
hThe column mean (SD) proportion represents the proportion of dermal contact with water duration that comes from other activities.
312
water. However, the participanting women were not different in terms of socioeconomic context or origin from the non- participants. Whereas they were older and more often employed than the non-participant women, age and employment have been described as associated only with filtered tap water ingestion during pregnancy, not with total tap water ingestion.4,10 Estimates of Water Ingestion
Tap water ingestion represented 71% of total water ingestion.
With similar definition of water ingestion and data collection procedures relatively similar to ours, Forsse´n et al5found 69% in United States of America, Smith et al10 78% in United Kingdom and Barbone et al223% in Italy. Bottled ingestion concerned 48%
of the EDDS women, as compared to 82% in Spain.7Proportion of tap water ingestion of EDDS women seemed to be comparable to that in English-speaking countries and lower than that in other Mediterranean countries, such as Italy or Spain.
Tap water ingestion of EDDS women is closer to one of the rare French studies on this topic that used the same estimation method as ours: 1.85±0.59 l/day for EDDS women versus 1.55±0.69 l/day for the general French population during winter and 1.89±0.59 l/day for EDDS womenversus1.78±0.84 l/day for the general French population during spring.24
A limit of our estimates is that we did not assess indirect tap water added for preparation in foods such as soup (95% of water), broth (70%) and noodles (70%) as did other authors.25,26Indirect tap water can represent 1.2–1.9 l/week, that is to say 23–25%
of total tap water ingestion.27Our results could consequently be underestimated.
Estimates of Dermal Contact with Water
Showering duration of EDDS pregnant French women (73 and 75 min/week, respectively, at second and third trimesters) was lesser than that of American pregnant women, who spend on Table 4. Analysis of factors associated with water ingestion of 132 pregnant women in the EDDS Cohort, Deux-Se`vres, France, 2012–2013.
Differences in mean ingestion of tap water (ml/day)
bunadjusted 95% CI P badjusted 95% CI P
Season at the second trimester of pregnancy 0.021 0.062
Winter Reference Reference
Spring 423 (36; 811) 384 (11; 780)
Summer 244 (627; 139) 244 (662; 174)
Autumn 58 (464; 348) 5 (414; 424)
Maternal age (years) 0.110 0.332
r31 Reference Reference
431 239 (55; 533) 161 (167; 490)
Education level 0.552 0.787
Primary-secondary school 232 (194; 658) 75 (582; 432)
French baccalaureate 134 (574; 305) 249 (738; 240)
Bacþ2 28 (338; 394) 105 (505; 295)
Higher education Reference Reference
Employment 0.225 0.385
No 242 (151; 636) 247 (808; 314)
Yes Reference Reference
BMIa 0.436 0.397
Underweight 47 (406; 500) 151 (600; 299)
Recommended weight Reference Reference
Overweight 313 (75; 702) 277 (157; 711)
Obese 182 (342; 707) 239 (296; 774)
Parity 0.049 0.395
0 Reference Reference
1 20 (363; 324) 15 (367; 397)
Z2 393 (18; 768) 275 (178; 728)
Smoking at the second trimester of pregnancy o103 0.013
Yes 968 (519; 1417) 777 (171; 1384)
No Reference Reference
Household income (h/month)b 0.694 0.719
801–1500 450 (618; 1519) 173 (188; 1540)
1501–2300 105 (624; 834) 285 (538; 1109)
2301–3000 353 (344; 1051) 498 (220; 1217)
3001–3800 101 (595; 797) 418 (288; 1125)
3801–4500 303 (508; 1114) 489 (318; 1296)
44500 Reference Reference
EPICES scorec 0.288 0.768
1: Less deprived Reference Reference
2 40 (372; 291) 50 (412; 312)
3 207 (337; 752) 17 (583; 549)
4 507 (208; 1222) 649 (395; 1693)
5: Most deprived 952 (254; 2159) 680 (1456; 2815)
aBMI: body mass index (underweight:o18.5 kg/m2; recommended weight: (18.5–25); overweight: (25–30); obeseZ30).
b1 Euro¼1.3645 US dollars.
cEPICES score is a French individual deprivation score.
Bold values are statistically significant relation at 5%.
313
average 105 min/week showering4but similar to that of pregnant Spanish or English women who spend on average 80 and 74 min/
week showering.7,10
Bathing duration of EDDS pregnant French women (33 and 34 min/week, respectively, at second and third trimesters) was less than that shown in other studies, which reported 46–87 min/week bathing4,5,8–10but greater than that for pregnant Spanish women.7 As the studies were not necessarily carried out during the same season and as they did not take any physical activities requiring additional showering or bathing into account, methodological differences could explain the divergent results.
Swimming duration of pregnant French women (59 and 54 min/
week) was moderate in comparison with other studies, in which estimates ranged all the way from 37to 156 min/week4but only 17–20% of women were involved, respectively, and seasonal differences could explain the wide variation.
As regards total dermal contact with water, the most widely found pathways were showering and hand-washing. However, hand-washing involves water contact with a smaller body area than showering, bathing or swimming. Furthermore, uptake of chemicals across the keratinized skin of palms probably differ from other body surfaces.
Table 5. Analysis of factors associated with mean shower duration of 132 pregnant women in the EDDS Cohort, Deux-Se`vres, France, 2012–2013.
Differences in mean duration of shower (min/week)
bunadjusted 95% CI P badjusted 95% CI P
Season at the second trimester of pregnancy 0.002 0.003
Winter Reference Reference
Spring 29 (12; 46) 30 (13; 48)
Summer 1(16; 18) 1 (19; 18)
Autumn 5 (23; 13) 0 (19; 18)
Maternal age (years) 0.110 0.509
r31 Reference Reference
431 11 (24; 2) 5 (19; 10)
Education level 0.534 0.924
Primary, secondary school 13 (7; 32) 1 (23; 22)
French baccalaureate 9 (11; 29) 5 (27; 16)
Bacþ2 1 (15; 18) 2 (15; 20)
Higher education Reference Reference
Employment 0.210 0.266
No 11 (6; 29) 14 (11; 39)
Yes Reference Reference
BMIa 0.008 0.033
Underweight 13 (33; 7) 17 (37; 3)
Recommended weight Reference Reference
Overweight 12 (5; 29) 7 (12; 26)
Obese 33 (10; 56) 26 (2; 50)
Parity 0.118 0.040
0 Reference Reference
1 7 (9; 22) 14 (3; 31)
Z2 11 (28; 6) 8 (28; 12)
Smoking at the second trimester of pregnancy 0.178 0.691
Yes 15 (7; 36) 5 (21; 32)
No Reference Reference
Household income (h/month)b 0.912 0.902
801–1500 8 (41; 56) 34 (110; 41)
1501–2300 5 (29; 38) 11 (48; 25)
2301–3000 5 (27; 36) 2 (34; 30)
3001–3800 4 (36; 27) 3 (34; 28)
3801–4500 5 (32; 41) 4 (32; 39)
44500 Reference Reference
EPICES scorec 0.259 0.428
1: Less deprived Reference Reference
2 15 (0; 30) 12 (4; 28)
3 6 (19; 30) 9 (16; 34)
4 24 (9; 56) 35 (12; 81)
5: Most deprived 21 (33; 76) 40 (54; 134)
aBMI: body mass index (underweight:o18.5 kg/m2; recommended weight (18.5–25); overweight: (25–30); obese(Z30).
b1 Euro¼1.3645 US dollars.
cEPICES score is a French individual deprivation score.
314
Differences between Trimesters
We found that total tap and bottled water ingestion did not change between second and third trimesters of pregnancy. Few studies have estimated water-use habits at two different stages of pregnancy.5,7 One found, likewise, an appreciable correlation between early and mid-pregnancy ingestion (r¼0.53 for tap water, 0.56 for bottled water and 0.40 for total water).5However, place of ingestion changed between second and third trimesters of pregnancy, as the result of work cessation.
The only difference found in dermal water contact between the second and the third trimesters pertained to hand-washing. There was no variation in showering. Indeed, intra-individual variations in this practice are known to be small.5Decrease in hand-washing duration between trimesters is mainly because of work cessation.
Estimating water-use habits during pregnancy should be carried out many times with a schedule depending on work cessation date.
Factors Associated with Water-use Habits
We found an effect of season on showering duration, as had been previously demonstrated,24,25,28 but not on tap water inges- tion, as indicated by Forsse´n et al.4
We did not find an effect of age on water-use habits during pregnancy, on contrary to Forsse´n et al4 who found more tap water ingestion in older women, and Smith et al10 who found a longer contact with water in younger women. As in previous studies,9,10 we did not find an effect of income, employment, educational level and EPICES score, but our results differed from those of Forsse´n et al, who found lower cold unfiltered tap water ingestion and higher cold filtered tap water ingestion in women with higher income than in women with lower income.4
Forsse´n et al4concluded that demographic variables were more strongly predictive of water-use habits during pregnancy than health and behavioral variables. However, we found that women who smoked during pregnancy had higher tap water ingestion, whereas Forsse´n et al showed it only for unfiltered tap water and bottled water. Moreover, we found an effect of BMI for showering duration, on contrary to Forsse´n et al.4Only one-third of northern inhabitants of France consume tap water, as opposed to 85% of those in the East and South of France. Therefore, French water ingestion seems to be not dependent on sociodemographic factors but rather on home location.29
These differences in water-use habits are important to take into account so as to avoid misclassification when estimating exposure, and confounding bias when studying the relationship between water exposure and health outcome.4
Strengths and Limitations of the Study
Many biases have been avoided by the EDDS study design.
Misunderstanding questions that could bias the reporting18 has been avoided through face-to-face interview by a trained inter- viewer, as described in 4 out of 10 previous studies (Supplemen- tary File 4). Memory bias of water exposure data4 has been avoided by a 1-day-recall interview. Diaries were used in four studies (Supplementary File 4) because they require little recall.18 However, using a 1-day-recall questionnaire as an exposure assessment method has been demonstrated to be the best alternative to diaries, especially when repeated at least once.30 Moreover, questionnaires and diaries had good agreement2,6 and the advantage of questionnaires is a higher response rate.9,18 Using a 1-day recall raises potential problems with respect to day of week and seasonal effects; however, we have scheduled home visit to take these point into account.
Information bias as concerns location of residence has been avoided by taking into account all places of ingestion. Indeed, at home, women drink the same volume of tap and bottled water, whereas at work they drink three times more bottled water than tap water.9
The limitations of the study include which are differences between participants and non-participants, low participation rate, and complete present data on only 132 of the 166 eligible women who agreed to participate. Generalizing from the participants to the overall EDDS cohort and eventually to the French population should be undertaken with the utmost caution.
CONCLUSION
French pregnant women from EDDS cohort study drink more tap water than bottled water and consumption was stable between second and third trimesters. Smoking is a predictor of tap water ingestion and season and obesity are the predictors of showering duration. These lines of information will help researchers to improve risk assessment of water contaminants.
CONFLICT OF INTEREST
The authors declare no conflict of interest.
ACKNOWLEDGEMENTS
We gratefully acknowledge the financial support of the CPER 2007–2013
‘‘Programme Eaux et Sols’’ (water-soil Programme) through local (Conseil Ge´ne´ral de la Vienne and Grand Poitiers), regional (Conseil Re´gional Poitou-Charentes), national (DRRT and Ministry of Education and Research) and European (FEDER) funds.
We wish to thank all the women who participated in the EDDS study. We wish to thank the Conseil Ge´ne´ral des Deux-Se`vres, service de protection maternelle et infantile for having shared their birth record data. We wish to thank Ce´line Girard (CG) and Fre´de´rike Limousi for recruitment; Julie, Marion and Fiona for data recording. We wish to thank Pascale Pierre-Euge`ne for laboratory assistance. We wish to thank the contribution of participating maternities: midwives team of Poitiers hospital, Dr. Pascal Villemonteix for Bressuire hospital, Dr. Claire Dekindt for Niort hospital, Dr. Christophe Re´gniez for Inkermann clinic, Dr. Xavier Aireau for Cholet hospital and their team of midwives and nurses. We wish to thank the Centre Maurice Halbwach and particularly Alexandre Kysch for having provided the correlation map between addresses and IRIS. Finally, we wish to thank Jeffrey Arsham, an American teacher and translator, for his English-language revision.
REFERENCES
1 Eskenazi B, Bradman A, Finkton D, Purwar M, Noble J, Pang Ret al.A rapid questionnaire assessment of environmental exposures to pregnant women in the INTERGROWTH-21(st) Project.BJOG2013;120(Suppl 2): 129–138.
2 Barbone F, Valent F, Brussi V, Tomasella L, Triassi M, Di Lieto Aet al.Assessing the exposure of pregnant women to drinking water disinfection byproducts.Epide- miology2002;13: 540–544.
3 Ershow AG, Brown LM, Cantor KP. Intake of tapwater and total water by pregnant and lactating women.Am J Public Health1991;81: 328–334.
4 Forssen UM, Herring AH, Savitz DA, Nieuwenhuijsen MJ, Murphy PA, Singer PC et al.Predictors of use and consumption of public drinking water among pre- gnant women.J Expo Sci Environ Epidemiol2007;17: 159–169.
5 Forssen UM, Wright JM, Herring AH, Savitz DA, Nieuwenhuijsen MJ, Murphy PA.
Variability and predictors of changes in water use during pregnancy.J Expo Sci Environ Epidemiol2009;19: 593–602.
6 Shimokura GH, Savitz DA, Symanski E. Assessment of water use for estimating exposure to tap water contaminants.Environ Health Perspect1998;106: 55–59.
7 Villanueva CM, Grimalt JO, Ballester F, Ibarluzea J, Sala M, Tardon A et al.
(Measurement of drinking water contaminants and water use activities during pregnancy in a cohort study in Spain).Gac Sanit2006;20(Suppl 3): 1–9.
8 Zender R, Bachand AM, Reif JS. Exposure to tap water during pregnancy.J Expo Anal Environ Epidemiol2001;11: 224–230.
9 Kaur S, Nieuwenhuijsen MJ, Ferrier H, Steer P. Exposure of pregnant women to tap water related activities.Occup Environ Med2004;61: 454–460.
10 Smith RB, Toledano MB, Wright J, Raynor P, Nieuwenhuijsen MJ. Tap water use amongst pregnant women in a multi-ethnic cohort.Environ Health2009;
8(Suppl 1): S7.
11 Costet N, Garlantezec R, Monfort C, Rouget F, Gagniere B, Chevrier C et al.
Environmental and urinary markers of prenatal exposure to drinking water dis- infection by-products, fetal growth, and duration of gestation in the PELAGIE birth cohort (Brittany, France, 2002-2006).Am J Epidemiol2012;175: 263–275.
315
12 Makris KC, Andra SS. Limited representation of drinking-water contaminants in pregnancy-birth cohorts.Sci Total Environ2013;468-469C: 165–175.
13 Dupuis A, Migeot V, Cariot A, Albouy-Llaty M, Legube B, Rabouan S. Quantification of bisphenol A, 353-nonylphenol and their chlorinated derivatives in drinking water treatment plants.Environ Sci Pollut Res Int2012;19: 4193–4205.
14 Qiao T, Yu Z, Zhang X, Au DW. Occurrence and fate of pharmaceuticals and personal care products in drinking water in southern China.J Environ Monitor 2011;13: 3097–3103.
15 Albouy-Llaty M, Dupuis A, Grignon C, Strezlec S, Pierre F, Rabouan S et al.
Estimating water intake in French pregnant women for bisphenol A risk assess- ment: the EDDS cohort study. Medicine faculty of Poitiers (ed). In: 6th Research day of Tours-Poitiers Hospital Universities. Poitiers, 2013, p 15.
16 Kuch HM, Ballschmiter K. Determination of endocrine-disrupting phenolic compounds and estrogens in surface and drinking water by HRGC-(NCI)-MS in the picogram per liter range.Environ Sci Technol2001;35: 3201–3206.
17 Vulliet E, Cren-Olive C. Screening of pharmaceuticals and hormones at the regional scale, in surface and groundwaters intended to human consumption.
Environ Pollut2011;159: 2929–2934.
18 Nieuwenhuijsen M. Introduction to exposure assessment. In: Nieuwenhuijsen M (ed). Exposure Assessement in Occupational and Environnmental Epidemiology.
Oxford University press: London, 2004, pp 4–18.
19 Bihan H, Laurent S, Sass C, Nguyen G, Huot C, Moulin JJet al.Association among individual deprivation, glycemic control, and diabetes complications: the EPICES score.Diabetes Care2005;28: 2680–2685.
20 Townsend P. Deprivation.J Soc Pol1987;16: 125–146.
21 INSEE. Correspondances adresses/IRIS 2000–1999. Centre Maurice Halbwachs (CMH). http://www.cmh.ens.fr. 18 February 2014.
22 ORS. Mother and child health in Poitou-Charentes 2013http://www.
insee.fr/fr/themes/detail.asp?reg_id=99&ref_id=etat-civil-naissances. 18 February 2014.
23 Ward LA, Cain OL, Mullally RA, Holliday KS, Wernham AG, Baillie PD et al.
Health beliefs about bottled water: a qualitative study.BMC Public Health2009;9:
196.
24 Gofti-Laroche L, Potelon JL, Da Silva E, Zmirou D. Description of drinking water intake in French communities (E.MI.R.A. study).Rev Epidemiol Sante Publique2001;
49: 411–422.
25 Ji K, Kim Y, Choi K. Water intake rate among the general Korean population.Sci Total Environ2010;408: 734–739.
26 Sebastian RS, Wilkinson Enns C, Goldman JD, Moshfegh AJ. Change in metho- dology for collection of drinking water intake in What We Eat in America/National Health and Nutrition Examination Survey: implications for analysis.Public Health Nutr2012;15: 1190–1195.
27 Beaudeau P, Zeghnoun A, Ledrans M, Volatier J. Consommation d‘eau du robinet pour la boisson en France me´tropolitaine: re´sultats tire´s de l‘enqueˆte alimentaire INCA1.Environnement Risques Sante´2003;2: 147–158.
28 Barraj L, Scrafford C, Lantz J, Daniels C, Mihlan G. Within-day drinking water consumption patterns: results from a drinking water consumption survey.J Expo Sci Environ Epidemiol2009;19: 382–395.
29 Menard C, Girard D, Leon C, Beck F (eds).Barome`tre sante´ environnement 2007, Paris, 2007.
30 Mons MN, van der Wielen JM, Blokker EJ, Sinclair MI, Hulshof KF, Dangendorf F et al.Estimation of the consumption of cold tap water for microbiological risk assessment: an overview of studies and statistical analysis of data.J Water Health 2007;5(Suppl 1): 151–170.
Supplementary Information accompanies the paper on the Journal of Exposure Science and Environmental Epidemiology website (http://
www.nature.com/jes) 316
